FITNESS CONSEQUENCES OF PARENTAL BEHAVIOR IN RELATION TO OFFSPRING NUMBER IN A PRECOCIAL

The Auk 111(3):563-572, 1994 FITNESS CONSEQUENCES OF PARENTAL BEHAVIOR IN RELATION TO OFFSPRING NUMBER IN A PRECOCIAL SPECIES: THE LESSER SNOW GOOSE TONY D. WILLIAMS, 1 MAARTEN J. J. E. LOONEN, 2 AND FRED COOKE 3 Department of Animal and Plant Sciences, P.O. Box 601, University of Sheffield, Sheffield S10 2UQ, United Kingdom; 2Zoological Laboratory, P.O. Box 14, University of Groningen, 9750 AA Haren, The Netherlands; 3Department of Biology, Queen's University, Kingston, Ontario K7L 3N6, Canada ABSTRACT.--We investigated the relationship between parental behavior and brood size, and the consequences of this relationship in terms of parental fitness (timing of molt and body mass at onset of molt in same year as breeding, and probability of return, timing of breeding, and clutch size in following year) in the precocial Lesser Snow Goose (Chen caerulescens caerulescens) at La P6rouse Bay, Manitoba. The percentage of time parent birds spent feeding decreased with increasing brood size, from greater than 90% for pairs without offspring to less than 80% for broods of seven and eight. The number of vigilant (head-up) postures per minute by parental birds increased up to brood size five and then decreased. Parental females also spent significantly less total time feeding and more time in alert behavior as brood size increased from one to five goslings. The relationship between parental behavior and brood size remained significant for small brood sizes even if pairs without goslings were excluded (range one to five goslings), and this relationship was independent of female age. Males (but not females) rearing larger broods molted later than those with smaller broods, although only by one to two days. This was directly related to rearing of offspring; both sexes, birds that hatched four or more goslings and subsequently lost one or more goslings during brood-rearing molted significantly earlier than birds rearing all of their hatched goslings. There was no relationship, in either sex, betwee number of goslings reared and the adult mass five to six weeks posthatch (molt) in the same year, or probability of return or timing of breeding (laying date or hatch date) in the following year. Partners of males that reared the largest number of goslings laid significantly larger clutch sizes the following year, suggesting that these were "better-quality" pairs. Over the range of naturally observed brood sizes, the effect of increasing brood size on parental behavior does not appear to be associated with any negativ effects on residual parental reproductiveffort or fitness in this species. Received 24 June 1993, accepted 18 November 1993. IN ALTRICIAL SPECIES, the evolutionary determination of clutch size and parental care are thought to be closely related; with increasing number of offspring there is a trade-off between benefits to the parent of increased inclusive fitness and costs of reduced residual reproductive value due to increased current reproductive effort (Perrins 1965, Williams 1966, Lack 1968, Trivers 1972, Drent and Daan 1980). The main form of parental care in altricial species is provision of food to the offspring. This form of parental care is considered to be "depreciable" (Altmann et al. 1977) or "shared" (Lazarus and Inglis 1986). The benefit of parental care per offspring declines with increasing number of offspring, and total parental expenditure is predicted to increase with increasing brood size (Lazarus and Inglis 1986). Numerous studies of altricial species have demonstrated costs of re- production (decreased fecundity and/or survival) associated with increased parental effort in rearing experimentally enlarged broods (e.g. Nur 1984, Hegner and Wingfield 1987, see review in Partridge 1990). In contrast to altricial species, the relationship between parental effort and number of off- spring in precocial species is much less clear (Schindler and Lamprecht 1987). Precocial species do not feed their young and the main form of parental care (vigilance or brood-defense be- havior) is generally assumed to be "nondepreciable" (Lazarus and Inglis 1986, Schindler and Lamprecht 1987), meaning that the total benefit of parental care is gained simultaneously by all offspring. In Southern Lapwings (Vanellus chilensis; Walters 1982), Bar-headed Geese (Anser indicus; Schindler and Lamprecht 1987), and Barnacle Geese (Branta leucopsis; Black and Owen 563

564 W LL ^MS, LOONEN, AND COOKE [Auk, Vol. 111 1989a, Forslund 1993), parents spent more time in vigilance behavior and less time foraging with increasing brood size, although no such relationship was found in Pink-footed Geese (A. bracyrhynchus; Lazurus and Inglis 1978) or in an earlier study of the Lesser Snow Goose (Chen caerulescens caerulescens; Lessells 1987). Schindler and Lamprecht (1987) suggested that in precocial species reduced foraging time of parent birds, associated with increasing brood size, may reduce parental fitness and, thus, represent a form of depreciable care. However, very few studies have considered actual fitness costs of band. Only female goslings were color banded because this specie shows highly female-biased philorearing different numbers of offspring in prepatty; therefore, only known-age females were available for behavioral observations. The presence of web cocial species. Parental care will only influence tags on goslings encountered at banding allowed the clutch size if the parent incurs such fitness costs, gosling's natal nest and parents to be identified, and in terms of decreased future fecundity and/or the number of goslings reared (brood size at banding, survival. Otherwise, optimal clutch size should BSB) to be determined. The timing of molt was debe infinite with respect to parental effort during termined for adults by measuring the length of the brood rearing (Lessells 1987). ninth primary at banding. To determine return rates In this paper we first analyze the relationship any record of resightings (during nesting or brood between parental behavior and brood size (de- rearing) and recaptures (during banding) of banded birds between 1973 and 1992 were used. As data were creased feeding performance and increased vigpooled over years, initial year of capture (the broodilance behavior) in the precocial Lesser Snow rearing year for which BSB was known) was restricted Goose. Second, as other parental traits (e.g. age to 1988 or earlier so that there were at least four or "quality") may be correlated both with the level of parental care and the number of offspring reared (Black and Owen 1989b, Clutton- Brock 1991), we investigate the effect of parental age on this relationship. Finally, correlational studies of parental behavior and brood size can be misleading because time (or energy) costs may not reflect fitness costs (Clutton-Brock 1991, Clutton-Brock and Godfray 1991). Therefore, we use data from a long-term population study of this species to investigate the relationship between number of offspring reared and five measures of residual parental fitness: (1) timing of molt and body mass at molt in the same year as brood rearing; and (2) probability of return, timing of breeding, and clutch size in the year following brood rearing. METHODS General methods and measures of parental fitness.--data on the breeding biology of the Lesser Snow Goose have been collected at La P rouse Bay, Manitoba (58ø43'N 93ø27'W) from 1968 to 1993. General field methods and details of the species' breeding biology are given in Cooke et al. (1985), Cooke and Rockwell (1988) and Cooch and Cooke (1991). In each year up to 2,000 nests were visited daily throughout hatch to obtain data on mean hatch date, the number of goslings leaving the nest (GLN), and the band numbers of breeding adults. All hatching goslings were marked with individually numbered web tags. For a smaller sample of nests (those found at the one-egg stage and visited daily throughout egg laying), the date of clutch initiation (laying date) and total clutch laid (TCL) were also known. Since 1972, 4,000 to 7,000 adults and goslings have been banded annually prior to fledging of the goslings, five to six weeks after mean hatch date. During banding all birds were sexed by cloacal examination, each received a metal Canadian Wildlife Service band, and female goslings and all adults received a year-specific colored alphanumeric potential re-encounter years (only 2% of birds were re-encountered five or more years after the broodrearing year, n = 3,466). Adult body mass and ninth primary length at banding varied significantly with year and absolute date. Therefore, data were converted to residuals from predicted year and date values (using the GLM procedure; SAS Institute 1990) before analysis of the relationship with respect to brood size. Similarly, there was significant annual variation in mean laying date, hatch date, and clutch size. Therefore, all data were converted to residuals from annual means before being pooled over years. All data were restricted to older birds (five or more years of age) for which age and fecundity are statistically independent in this population. Data on foraging behavior were collected in 1985, 1986, and 1992, and general activity budgets were recorded in 1991 and 1992 during the brood-rearing period. All observations were made between hatch and banding from a permanent, 5-m-high tower situated on intertidal feeding flats (Randy's Fiats) adjacent to the nesting area. For all pairs from which data were obtained, at least one adult was banded (and, therefore, of known sex) and, whenever possible, pairs with known-age females were chosen. Ob- servations were made throughout daylight hours, from 0500-2200 CST, with a minimum of 24 records obtained for each hourly interval. Due to disturbance from other researchers on the feeding flats, more observations were made in the early morning (0500-0900, 48%) than during the later part of the day, but

July 1994] Fitness Consequences of Parental Behavior 565 time of observation was not biased with respect to brood size. Foraging behavior.--foraging behavior was recorded continuously. Observations started once the bird adopted the head-down feeding posture and ended after 5 (1992) or 10 (1985, 1986) min, or when the bird changed to nonforaging behavior (preening, sleeping, or drinking). The following were recorded: (1) total observation time; (2) total time spent feeding (in head-down position); (3) number of feeding bouts, delimited by the bird stopping feeding and adopting the head-up posture while either stationary or moving; and (4) number of steps during feeding bouts. In addition, peck rates were measured at the beginning or end of each observation by counting the time taken for 50 pecks during bouts of continuous foraging (i.e. with no interruptions by head-up postures; Sedinger and Raveling 1988). Using these data the following additional variables were calculated: (1) number of pecks per minute during continuous feeding; (2) percentage of time spent feeding (feeding time/total study time); (3) number of vigilant, headup postures per minute; and (4) number of steps per minute feeding. Sex, brood size, and band information also were recorded for all individuals. 95 n:25 114 157 165 106 48 7 3 2 We assumed that the proportion of total 5-s activity record spent in each activity reflected the actual pro- Foraging behavior in relation to brood size was portion of time spent in each activity (following Eberanalyzed by multiple regression using the GLIM stahardtet al. 1989). Contingency table analysis and tistical package (Numerical Algorithms Group 1987), Spearman rank correlations were used to analyze accorrecting for variation in adult sex, days since mean tivity budgets in relation to sex, year, brood size, or hatch, and year. Dependent variables were transfemale age. All analyses were carried out using SAS formed to improve the distribution of the residuals: (SAS Institute 1990). percentage time feeding (arcsine transformed), number of head-ups, and steps per minute feeding (both square-rootransformed). Models were constructed RESULTS using a stepwise procedure including both linear and quadratic terms for the independent variables. Tests Foraging behavior.--the percentage time spent for inclusion of independent variables were based on feeding by parental birds decreased linearly an F-distribution of the change in deviance per de- with increasing brood size (Fig. 1; F = 15.79, df gree of freedom, when the variable was dropped from = 1 and 608, P < 0.001) from greater than 90% the model, divided by the deviance per degree of for zero brood sizes to less than 80% for brood freedom in the full model. Brood size, days since sizes of seven and eight. The number of alert mean hatch date, year, and sex were included as in- (head-up) postures per min by parents also vardependent variables, and year and sex were entered as factors. ied significantly with brood size (Fig. 2; F = Activity budgets.--activity budgets were recorded, 12.41, df = 2 and 617, P < 0.001), increasing for brood sizes from one to five and then decreasin addition to detailed feeding behavior, to test the possibility that differences in foraging behavior in ing for larger broods. The relationships berelation to brood size might be compensated for by tween these two behavioral variables and brood changes in other behaviors. These were obtained us- size were not significantly different in male and ing a scan-sampling technique, activity being record- female birds (ANCOVA, sex x brood interaced at 5-s intervals over 4-min periods (n = 48 records). tion; feeding, F = 1.78; head-ups, F = 0.12). In all cases activity of both birds of a pair was recorded There was no significant relationship between consecutively. The following activities were distinnumber of pecks or number of steps per minute guished: feeding (head-down), alert or vigilant (headfeeding and brood size (P > 0.05). Trends in up plus extreme head-up of Lazarus and Inglis 1978), motor (swimming or walking, and not feeding), com- parental behavior with brood size have often fort (preening, stretching or resting), drinking, and been reported to be dependent on inclusion of social (usually inter- or intrabrood interactions). Birds pairs without offspring (e.g. Sedinger and Ravbrooding goslings were excluded from observations. eling 1990). Therefore, data were reanalyzed for -- 80 7ø f 65 0 1 2 3 4 5 6 7 8 9 brood size FIG. l. Relationship between percentage of time spent feeding and brood size in Lesser Snow Geese. œ + SE, with sample sizes above points. Equation for line is Y = 1.237-0.0202X, where X is brood size.

566 W LLL MS, LOONEN, AND COOKE [Auk, Vol. 111 4,5 n=25 114 157 165 106 48 7 3 2 TABLE 2. Activity budgets of male and female Lesser Snow Geese in relation to brood size. 4.G 3.5 3.G 1.5 1.0 0.5 0 1 2 3 4 5 6 7 8 9 brood size F G. 2. Relationship between number of head-up postures per minute and brood size in Lesser Snow Geese. œ + SE, with sample sizes above points. Equation for line is Y = 1.050 + 0.1418X - 0.0156X 2, where X is brood size. brood sizes of one to five only. For this restricted data set, both behavioral parameters were still significantly related to brood size (percentage time spent feeding, F = 7.65, df = 1 and 575, P < 0.01, number of head-ups per minute, F = 3.72, df = 2 and 580, P < 0.05). Activity budgets.--there were significant differences in activity budgets between sexes (X 2 = 246.8, df = 5, P < 0.001, data pooled across years) so males and females were analyzed separately. Males spent more time alert and less time feeding than females (Table 1). The proportion of time spent in different behaviors did not vary between years in either sex (males, X 2 = 0.215; females, X 2 = 0.616; comfort, drinking, and social behaviors accounted for less than 5% of the total time, so these behaviors were pooled Brood Percent time spent in each activity size n Feeding Alert Motor Other Female 1 37 85.2 3.3 6.4 5.1 2 64 87.8 6.0 4.8 1.4 3 72 83.8 5.2 5.8 5.2 4 48 87.9 5.2 5.3 1.6 5 12 75.7 8.5 8.5 7.3 Male 1 37 81.2 8.6 6.6 3.6 2 64 82.9 10.6 5.7 0.9 3 72 78.9 11.3 6.3 3.5 4 48 79.5 12.2 6.0 2.2 5 12 79.2 10.4 8.7 1.7 with motor behavior). Therefore, data were pooled across years. In females, for brood sizes of one or more, the proportion of total time spent alert and time spent feeding were positively (r = 0.14, n = 239, P < 0.05) and negatively (r = -0.11, n = 239, P = 0.07) correlated with brood size, respectively (Table 2). Time spent in motor or other activities was not related to brood size (P > 0.50). Females without goslings spent significantly more time feeding and less time in alert and other behaviors than females with one or more goslings (Table 1; X 2 = 49.8, df = 2, P < 0.001). With pairs without offspring included, the time spent alert and time spent feeding were more highly positively (r = 0.22, n = 253, P < 0.001) and negatively (r = -0.18, n = 253, P < 0.01) correlated with brood size, respectively. For individuals with goslings, the proportion of time spent alert was negatively correlated with time spent feeding (r = -0.535, P < 0.0001), TABLE 1. Activity budgets of male and female Lesser Snow Geese with and without broods. Values are percentages of total number of observations. Activity Year Sex n Feeding Alert Comfort Motor Drink Social With brood 1991 M 147 80.1 10.7 1.6 6.2 1.2 0.2 1991 F 147 84.2 5.6 3.1 5.5 1.5 0.1 1992 M 92 81.5 11.1 0.3 6.0 1.1 <0.1 1992 F 92 87.4 5.5 0.4 5.6 1.0 <0.1 Without brood 1992 M 14 92.3 2.2 0.0 5.5 0.0 0.0 1992 F 14 94.9 0.9 0.0 4.0 0.2 0.0

July 1994] Fitness Consequences of Parental Behavior 567 TABLE 3. Adult mass and timing of molt (indicated by ninth-primary length) in relation to brood size at banding (BSB, 5-6 weeks posthatch), and timing of molt in relation to brood loss (number of goslings reared compared to number leaving the nest, GLN-BSB). Values are mean residuals (to correct for variation with year and date), with standard error and sample size. BSB Adult mass (g) Ninth-primary length (mm) _+_+ SE (n) BSB œ + SE (n) Loss œ _+ SE (n) Females 0-7.8 + 10.9 (93) 0 3.25 -+-+ 2.92 (64) 0-1.55 + 2.86 (73) 1 12.7 + I1.2 (151) 1 1.23 -+-+ 2.50 (108) 1-2.33 -+ 2.36 (109) 2-4.4 + 8.6 (230) 2 0.86 -+-+ 1.78 (135) 2 2.34 + 2.55 (65) 3-7.8 + 6.7 (303) 3-1.83 + 1.79 (172) 3 2.07 -+-+ 3.88 (38) 4 8.5 + 9.7 (207) 4-2.14 -+-+ 2.39 (111) 4 5.78 + 4.17 (25) 5 5.1 -+-+ 12.7 (85) 5 3.05 + 3.87 (31) Males 0 9.3 + 26.6 (24) 0 15.53 + 4.55 (15) 0-3.49 _+ 3.76 (39) 1 17.1 + 26.6 (48) 1 10.19 _+_+ 4.27 (35) 1-4.80 _+_+ 2.77 (45) 2-12.1 + 16.9 (96) 2-1.33 -+-+ 2.80 (60) 2-4.75 + 4.35 (34) 3 4.6 + 11.6 (133) 3-2.99 + 2.75 (71) 3 7.37 _+_+ 5.59 (16) 4-9.5 + 15.0 (107) 4-4.12 + 2.37 (63) 4 12.61 _+ 5.13 (12) 5 11.7 _+_+ 24.0 (45) 5-2.92 + 7.88 (13) but not with time spent in other behaviors (P > 0.30). In males, for brood sizes of one or more, there was no correlation between brood size and either proportion of total time spent alert or time spent feeding (Spearman rank, r = 0.09, n = 239 and r = -0.06, n = 239, respectively; Table 2). Similarly, time spent in motor or other activities was not related to brood size (P > 0.50). However, as for females, males without goslings spent significantly more time feeding and less time in alert and other behaviors than males with one or more goslings (Table 1; X 2 = 62.7, df = 2, P < 0.001). If these pairs without offspring were included, time spent alert and time spent feeding were significantly positively (r = 0.18, n = 253, P < 0.01) and negatively (r = -0.14, n = 253, P < 0.05) correlated with brood size, respectively. For individuals with goslings, the proportion of time spent alert was negatively correlated with time spent feeding (r = -0.658, n = 239, P < 0.0001), but not with time spent in other behaviors (P > 0.20). Effect of female age on behavior.--for knownage females, using data on foraging behavior, there was no significant relationship between female age and either percentage time spent feeding (F = 0.07) or number of head-ups per minute (F = 0.13). If the foraging-behavior data were restricted only to known-age females, and female age was included as a variable in the model, both the behavioral variables were still significantly related to brood size (percentage time spent feeding, F = 5.19, df = 1,263, P < 0.05; number of head-ups per minute, F = 9.56, df = 1,269, P < 0.01). Parental fitness and brood size.--adult body mass at banding did not vary significantly with number of offspring reared (BSB) in either males or females, either with pairs without offspring included (Table 3) or excluded (P > 0.70 in all cases). Similarly, adult mass was not related to brood loss (GLN-BSB) in either sex (P > 0.70 in all cases). There was no significant difference in the number of days between hatch and banding for different brood sizes (BSB), so we did not explicitly control for this in subsequent analyses. Ninth-primary length decreased significantly with increasing BSB (for all brood sizes) in males (F = 13.41, P < 0.001) by an average 3.96 mm for each additional gosling (i.e. males with larger broods molted significantly later; Table 3). Primary length also decreased in females, by 0.91 mm per additional gosling, but this was not significant (F = 1.74; Table 3). This relationship remained significant in males if data were restricted to broods of one or more (3.51 mm per gosling; F = 7.65, P = 0.006). For birds hatching four or more goslings, ninth-primary length increased significantly with increasing brood loss in males (4.01 mm per gosling lost, F = 6.28, P = 0.013), and this relationship approached significance in females (2.27 mm per

568 WILLIAMS, LOONEN, AND CooszE [Auk, Vol. 111 TABLE 4. Effect of brood size at banding (BSB) on laying date, hatch date, clutch size, and probability of being re-encountered in the following year. Laying date, hatch date, and clutch size given as residuals from annual means and pooled over years; œ _+ SE (n). Percent Laying date Hatch date Clutch size re-encountered a 1 0.47 _+ 0.39 (17) 2-0.19 _+ 0.45 (18) 3-0.44 ñ 0.37 (27) 4-0.03 _+ 0.57 (16) 5 b -0.71 _+ 0.80 (8) 6 b _ 1 2 3 4 5 b Females -0.18 +_ 0.28 (62) 0.17 _+ 0.14 (62) 68.1 (160) -0.27 _+ 0.21 (93) 0.09 _+ 0.12 (93) 68.5 (276) -0.23 + 0.19 (78) 0.31 + 0.14 (78) 65.0 (257) -0.50 +_ 0.27 (56) 0.56 + 0.17 (56) 64.6 (192) 0.35 _+ 0.40 (28) 0.27 +_ 0.19 (28) 73.0 (74) -- -- 83.3 (12) Males -0.84 _+ 0.43 (24) 0.00 + 0.80 (24) 68.1 (47) -0.32 + 0.37 (33) 0.08 + 0.17 (33) 71.7 (92) -0.20 + 0.30 (29) 0.17 + 0.24 (29) 64.1 (92) -0.30 _+ 0.40 (28) 0.80 _+ 0.23 (28) 70.4 (81) 0.16 + 0.72 (8) 0.52 + 0.33 (8) 80.5 (41) ß 'n in parentheses. Brood sizes 5 and 6 pooled for laying date, hatch date, and clutch size, as well as for return rate in males. gosling lost, F = 3.23, P = 0.071; Table 3); birds that had the greatest reduction in brood size molted significantly earlier. There was no significant difference in laying date (in females), or hatch date (in both sexes) in the following year, relative to the number of offspring reared the previous year (P > 0.20 in all cases; Table 4). Residual mean clutch size increased with number of goslings reared the previous year in males (b = 0.22, F = 7.12, P = 0.009); males rearing larger broods bred with females that laid larger clutches the following year. The relationship between clutch size and number of goslings reared was also positive in females (b = 0.09), but this relationship was not significant (F = 2.90, P = 0.09; Table 4). The probability of birds being re-encountered in any subsequent year, up to four years after BSB, did not vary with number of goslings (BSB) reared in either males (X 2 = 3.15, df = 4, P > 0.50) or females (X 2 = 6.10, df = 5, P > 0.30; Table 4). Similarly, among those birds which were subsequently reencountered there was no difference in the probability of the bird being seen in the year following BSB compared to two or more years later, relative to BSB (males, X 2 = 3.79, df = 4, P > 0.40, females, X 2 = 3.98, df = 5, P > 0.50); that is, there was no evidence that birds rearing larger broods were more likely to skip a breeding attempt. DISCUSSION Male and female Lesser Snow Geese had significantly different overall activity budgets dur- ing brood rearing at La P6rouse Bay; males spent more time alert and less time feeding compared to females (although both sexeshowed the same trends in behavior with respect to brood size). This sex difference in behavior clearly was related to the presence or absence of offspring. In pairs without goslings, activity budgets did not differ between sexes. Several other studies sim- ilarly have shown that female geese spend more time feeding than males during brood rearing (Harwood 1977, Lazarus and Inglis 1978, Lessells 1987) and that males spend more time in the alert, head-up posture (Black and Owen 1989a, b, Sedinger and Raveling 1990). Sedinger and Raveling (1990) suggested that male Canada Geese (Branta canadensis) perform more brood defense and vigilance behavior so that their mates can spend more time feeding in order to recover body reserves lost during egg laying and incubation (Raveling 1979). Male Lesser Snow Geese spent more time in alert (vigilance) behavior than did females for each brood size (Table 1). This suggests that the role of male Lesser Snow Geese may actually be more importanthan that of females in brood defense and vigilance behavior. Martin et al. (1985) reported a rapid reduction in brood size of single parent females, and decreased gosling survival (in one year) in broods where the male was experimentally removed, compared to biparen- tal families. Foraging behavior of Lesser Snow Geese was significantly related to brood size in both sexes; birds adopted the alert (vigilant) posture more frequently with increasing brood size up to five

July 1994] Fitness Consequences of Parental Behavior 569 goslings and spent less time feeding with increasing brood size up to seven to eight goslings. We did not manipulate brood size to create experimentally enlarged broods; however, 99.8% (n = 3,466) of broods encountered at banding comprised six or fewer goslings. In addition, the relationship between parental behavior and number of offspring existed over a limited range of smaller brood sizes (from one to five goslings). Data from overall activity budgets also were consistent with this result for females, although not significantly so for males. Females spent a greater proportion of total time alert and a lesser proportion of time feeding as brood size increased from one to five goslings. In males, although activity budgets did not vary significantly with number of goslings, the trends for male behavior and brood size were in the same direction as those for females. Rockwell et al. (1993) showed that older Lesser Snow Geese (seven or more years of age) had lower repro- ductive success and that this was partly due to increased rates of total brood failure. However, we found no difference in mean female age relative to brood size in our sample. In addition, among known-age birds there was still a significant positive relationship between vigilance behavior and brood size, and a significant negative relationship between feeding behavior and brood size, controlling for female age. Schindler and Lamprecht (1987) similarly concluded that variation in parental behavior with increasing brood size was independent of female age in Bar-headed Geese. In females (and in males considering all in- dividuals), there was an inverse relationship between time spent alert and time spent feeding, but there was no systematichange in the amount of time spent in other activities. Increases in the time spent in vigilance behavior, with increasing brood size, therefore, are not compensated for by changes in other nonfeeding behaviors. Movement (walking and swimming) and drinking comprised between 70 and 95% of these other behaviors in male and female Lesser Snow Geese, respectively. Continuous movement is an integral part of the foraging strategy of geese because birds rapidly deplete food patches (Prop and Loonen 1988) and have to find new areas to graze. Regular drinking may also be essential in order to maintain water balance in this species which forages mainly on salt-marsh plants. Lesser Snow Geese, therefore, may be unable to further reduce the amount of time they spend in these other activities, even though this would allow an increase in vigilance behavior while maintaining time spent feeding. In contrasto the present study, Lessells (1987) found no relationship between time spent feeding and brood size in Lesser Snow Geese at La Pbrouse Bay in an earlier year (1979). This dif- ference may reflect the long-term decrease in food availability (by up to 75%) that has occurred at La P rouse Bay over the last decade (Williams et al. 1993; although there were also some methodological and analytical differences between the two studies; C. M. Lessells pers. comm.). Parental behavior may now be constrained by food availability and the adults own food requirements (Loonen et al. in prep.). Schindler and Lamprecht (1987) also found an increase in vigilance behavior and a decrease in time spent feeding in female Bar-headed Geese, but they found no such relationship in males. In contrast, for Canada Geese the percentage of foraging periods spent in the alert posture was greater in males than females, and the proportional increase in vigilance behavior with increasing brood size was twofold greater in males. If pairs without offspring were excluded, time spent alert increased with brood size in males but not in females (Sedinger and Raveling 1990). Several studies of precocial species, therefore, have now demonstrated a reduction in time spent feeding by parental birds, usually as a consequence of increased vigilance behavior, with increasing brood size (see also Madsen 1981, Waiters 1982, Schindler and Lamprecht 1987, Forslund 1993). Schindler and Lamprecht (1987) suggested that reduced feeding time in female Bar-headed Geese, with increasing brood size, might induce a slower, or incomplete, recovery of body condition following incubation, thus reducing fe- male fitness. However, the only studies to date that have investigated fitness consequences of brood size in precocial species are those of Roh~ wer (1985), Lessells (1986), and Rohwer and Heusmann (1991). Female Canada Geese that reared experimentally enlarged broods had lower body mass at the end of the brood-rearing period, molted later, and laid eggs later the following year, compared to those with smaller broods. However, there was no effect of brood size on overwinter survival or on clutch size in the following year (Lessells! 986). Rohwer (1985) found no relationship between female mass at the end of the rearing period and brood size in Blue-winged Teals (Anas discors), and adult fe-

570 WILLIAMS, LOONEN, AND COOK [Auk, Vol. 111 male survival was not related to brood size in the Wood Duck (Aix sponsa; Rohwer and Heusmann 1991). In Lesser Snow Geese, there was no correlation between brood size and adult mass five to six weeks posthatch, but males rearing larger broods molted later than those with small broods. The fact that birds (of both sexes) that lost goslings during brood rearing had significantly more advanced molt, compared to birds losing no goslings, suggests that delayed molt was directly related to the rearing of large broods. Forslund (1993) showed that in Barnacle Geese parents responded immediately to a decrease in brood size by decreasing their vigilance time. Delayed onset of, or slower, molt may be disadvantageous in arctic-nesting geese because it reduces the length of the final feeding phase, between completion of molt and departure from the breeding grounds. This period may be critical in achieving the condition nec- essary to successfully complete migration (Prop et al. 1984, Owen and Black 1989). Nevertheless, male Lesser Snow Geese with brood sizes of five had ninth primaries that on average were only 13 mm shorter than those with brood sizes of one. As primary feathers grow at a rate of 7 to 8 mm per day in Lesser Snow Geese (Hanson and Jones 1976), the delay in completion of molt would only have been of the order of one to two days. This relatively small effect is consistent with the fact that we found no evidence for a negative effect of increasing brood size on the probability of return in the following year or any residual effect on timing of breeding the following year. In contrast, birds that reared the largest number of goslings laid significantly larger clutches the following year, suggesting that birds rearing the largest number of goslings were, in fact, "better-quality" birds. This is supported by the fact that the trends in return rate and timing of breeding with brood size also were positive rather than negative. Several pre- vious studies have suggested that larger brood sizes may actually be advantageous in geese (Lamprecht 1986, Gregoire and Ankney 1990, Williams in press). For example, in most species, family groups are maintained during the winter. Offspring contribute increasingly to vigilance behavior and assist in conflicts with neighbors throughout the winter, allowing an increase in adult feeding time (Black and Owen 1989a, b). Thus, large brood sizes may enhance, rather than decrease, overwinter survival. Similarly, Prop et al. (1984) showed that larger broods were more dominant and tended to feed in the areas of highest food availability on the breeding grounds. In conclusion, parental behavior was significantly related to the number of offspring reared in Lesser Snow Geese at La P rouse Bay; parents of larger broods spent less time feeding and more time in vigilance behavior. However, over the range of naturally observed brood sizes we found no evidence that this was negatively related to parental fitness or future reproductive effort in this precocial species. We are aware of the difficulties of detecting true costs of reproduction (sensu Lessells 1991); additional studies of parent-offspring behavior are required for precocial species, but these should measure both changes in parental behavior and the consequences of these changes to parental fitness. ACKNOWLEDGMENTS We thank all the people who helped with tower observations and data collection during this study, and the many people who have contributed to the LPB Snow Goose project since 1976. Kate Lessells provided extensive comments on an earlier draft of this manuscript, and Rudi Drendt and Matthew Collins also provided useful comments. Lesser Snow Goose research at La P rouse Bay has been supported by the Natural Sciences and Engineering Research Council of Canada (NSERC), the Manitoba Department of Renewable Resources, the Wildlife Management Institute, the Canadian Wildlife Service, the Mississippi and Central Flyway councils, Ducks Unlimited and Ducks Unlimited Incorporated. T.D.W. was supported by an NSERC International Research Fellowship. M.J.J.E.L. received additional support from the Netherlands Foundation for Scientific Research in the Arc- tic. LITERATURE CITED ALTMANN, $. A., $. $. WAGNER, AND $. LENNINGTON. 1977. Two models for the evolution of polygyny. Behav. Ecol. Sociobiol. 2:397-410. BLACK, J. M., AND M. OwEN. 1989a. Agonistic behavior in Barnacle Goose flocks: Assessment, investment and reproductive success. Anim. Behav. 37:199-209. BLACK, J. M., AND M. OwEN. 1989b. Parent-offspring relationships in wintering Barnacle Geese. Anim. Behav. 37:187-198. CLUTTON-BROCK, T. H. 1991. The evolution of parental care. Princeton Univ. Press, Princeton, New Jersey. CLUTTON-BRocK, T. H., AND C. GODFRAY. 1991. Parental investment. Pages 234-262 in Behavioral ecology: An evolutionary approach (J. R. Krebs

July 1994] Fitness Consequences of Parental Behavior 571 and N. B. Davies, Eds.). Blackwell Scientific Publications, Oxford. COOCH, E.G., A D F. COOKE. 1991. Demographic changes in a snow goose population: Biological and demographic implications. Pages 168-189 in Bird population studies (C. M. Perrins, J-D. Lebreton, and G. J. M. Hirons, Eds.). Oxford Univ. Press, Oxford. COOKE, F., C. S. FINDLAY, R. F. ROCKWELL, AND J. A. SMITH. 1985. Life-histories of the Lesser Snow Goose (Anser caerulescens). III. The selective value of plumage polymorphisms: Net fecundity. Evolution 39:165-177. COOKE, F., AND R. F. ROCKWELL. 1988. Reproductive success in a Lesser Snow Goose population. Pages 237-250 in Reproductive success (T. H. Clutton- Brock, Ed.). Univ. Chicago Press, Chicago. DREhrr, R. H., AND S. D - N. 1980. The prudent parent: Energetic adjustments in avian breeding. Ardea 68:225-252. EBERHARDT, L. E., G. G. BROOKS, R. G. ANTHONY, AND W. H. RICKHARD. 1989. Activity budgets of Canada Geese during brood rearing. Auk 106:218-224. FORSLUND, P. 1993. Vigilance in relation to brood size and predator abundance in the Barnacle Goose Branta leucopsis. Anim. Behar. 45:965-973. GREGOIRE, P. E., AND C. D. ANKNEY. 1990. Agonistic behavior and dominance relationships among Lesser Snow Geese during winter and spring migration. Auk 107:550-560. HANSON, H. C., AND R. L. JONES. 1976. The biogeochemistry of Blue, Snow and Ross' geese. Southern Illinois Univ. Press, Carbondale. HARWOOD, J. 1977. Summer feeding ecology of Lesser Snow Geese. J. Wildl. Manage. 41:48-55. HEGNER, R. E., AND J. C. WINGFIELD. 1987. Effects of brood-size manipulations on parental investment, breeding success, and reproductive endocrinology of House Sparrows. Auk 104:470-480. LACK, D. 1968. Ecological adaptations for breeding in birds. Chapman and Hall, London. L MPRECvrr, J. 1986. Structure and causation of the dominance hierarchy in a flock of Bar-headed Geese (Anser indicus). Behavior 96:28-48. LAZARUS, J., AND I. R. INGLIS. 1978. The breeding behavior of the Pink-looted Goose: Parental care and vigilant behavior during the fledging period. Behavior 65:62-88. LAZARUS, J., AND I. R. INGLIS. 1986. Shared and unshared parental investment, parent-offspring conflict and brood size. Anim. Behar. 34:1791-1804. LESSELLS, C. M. 1986. Brood size in Canada Geese: A manipulation experiment. J. Anim. Ecol. 55: 669-689. LESSEEES, C. M. 1987. Parental investment, brood size and time budgets: Behavior of Lesser Snow Goose families. Ardea 75:189-203. LESSELLS, C.M. 1991. The evolution of life histories. Pages 32-68 in Behavioural ecology: An evolutionary approach (J. R. Krebs and N. B. Davies, Eds.). Blackwell Scientific Publications, London. MADSEN, J. 1981. Post-hatching behavior of families and non-breeding Greenland White-fronted Geese. Pages 116-122 in Report of the 1979 Greenland White-fronted Goose Study Expedition to Edalungmiut Nunat, West Greenland (A.D. Fox and D. A. Stroud, Eds.). Aberystwyth, Wales. MARTIN, K., F. G. COOCH, R. F. ROCKWELL, AND F. COOKE. 1985. Reproductive performance in Lesser Snow Geese: Are two parents essential? Behav. Ecol. Sociobiol. 17:257-263. NUMERICAL ALGORITHMS GROUP. 1987. GLIMmanual release 3.77. Numerical Algorithms Group, Oxford. NUR, N. 1984. The consequences of brood size for breeding Blue Tits: I. Adult survival, weight change and the cost of reproduction. J. Anim. Ecol. 53:479-496. OWEN, M., AND J. M. BLACK. 1989. Factors affecting the survival of Barnacle Geese on migration from the breeding grounds. J. Anim. Ecol. 58:603-617. PARTRIDGE, L. 1990. Lifetime reproductive success and life-history evolution. Pages 422-440 in Lifetime reproduction in birds (I. Newton, Ed.). Academic Press, London. PERRINS, C. M. 1965. Population fluctuations and clutch size in the Great Tit Parus major L. J. Anim. Ecol. 34:601-647. PROP, J., AND M. J. J. E. LOONEN. 1988. Goose flocks and food exploitation: The importance of being first. Pages 1878-1887 in Acta XIX Congressus Internationalis Ornithologici (H. Ouellet, Ed.). Ottawa, Ontario, 1986. National Museum of Natural Sciences, Ottawa. PROP, J., M. R. VAN EERDEN, AND R. H. DRENT. 1984. Reproductive success of the Barnacle Goose Bran- ta leucopsis in relation to food exploitation on the breeding grounds, western Spitzbergen. Norsk Polar. Skrift. 181:87-117. RAVELING, D. G. 1979. The annual cycle of body composition of Canada Geese with special reference to control of reproduction. Auk 96:234-252. ROCKWELL, R. F., E.G. COOCH, C. B. THOMPSON, AND F. COOKE. 1993. Age and reproductive success in female Lesser Snow Geese: Experience, senescence and the cost of philopatry. J. Anim. Ecol. 62:323-333. ROI-IWER, F. C. 1985. The adaptive significance of clutch size in prairie ducks. Auk 102:354-361. ROHWER, F. C., AND H. W. HEUSMANN. 1991. Effects of brood size and age on survival of female Wood Ducks. Condor 93:817-824. SASINSTITtYrE. 1990. SAS user's guide: Statistics, version 6. SAS Institute, Inc., Cary, North Carolina. SCHINDLER, M., AND J. LAMPRECHT. 1987. Increase of

572 WILLIreS, LOONEN, qi> COO <E [Auk, Vol. 111 parental effort with brood size in a nidifugous bird. Auk 104:688-693. SEDINGER, J. S., AND D. G. RAVELING. 1988. Foraging behavior of cackling Canada Goose goslings: Implications for the roles of food availability and processing rate. Oecologia 75:119-124. SEDINGER, J. S., AND D. G. RAVELING. 1990. Parental behavior of Cackling Canada Geese during brood rearing: Division of labor within pairs. Condor 92:174-181. TRIVERS, R. L. 1972. Parental investment and sexual selection. Pages 136-179 in Sexual selection and the descent of man 1871-1891 (B. Campbell, Ed.). Aldine Press, Chicago. WALTERS, J. R. 1982. Parental behavior in Lapwings (Charadriidae) and its relationship with clutch size and mating systems. Evolution 36:1030-1040. WILLIAMS, G.C. 1966. Natural selection, the costs of reproduction, and a refinement of Lack's principle. Am. Nat. 100:687-690. WILrlAMS, T. D. 1994. Adoption behavior in a precocial species, the Lesser Snow Goose: Intergenerational conflict, altruism or a mutually-beneficial strategy? Anim. Behav. In press. WILLIAMS, T. D., E.G. COOCH, R. L. JEFFERIES, AND F. COOICE. 1993. Environmentaldegradation, food limitation and reproductive output: Juvenile survival in Lesser Snow Geese. J. Anim. Ecol. 62: 766-777.